Discrimination, Segmentation and Vulnerability in the Labor Market Luca Flabbi
Why Market Failures in the Labor Market? Because the good traded in labor markets (Labor: supplied by workers and demanded by firms) has peculiar characteristics: Labor is NOT: as homogenous as tradable as any other good
Why Market Failures in the Labor Market? (cont.) Examples: Workers productivity is heterogeneous and (often) unobservable Workers and employers (or jobs ) mobility is costly, generating frictions Productivity is the result of a matching process: the same worker at a different firm may generate different productivity
Consequences of Labor Market Failures Labor Market Segmentation Inefficient allocation of resources, generating rents Survival of Discrimination Discrimination is not the result of market failures but its survival and persistency is Again, inefficient allocation of resources
Discrimination: Definition Definition: Labor market discrimination is defined as the systematically different treatment reserved to members of a given group, just because they belong to that group The result is a distortion in prices and quantities, i.e. in wages and in workers allocation to jobs Example: wage discrimination is defined as members of a given group receiving a lower wage at same productivity just because they belong to that group
Discrimination: Definition (cont.) Groups against which discrimination is exerted are usually vulnerable groups: Women Young or Old Belonging to Minority Ethnic Groups Members of a vulnerable group are: Weak and liable to serious hardship Or face an uninsured risk to end up in this condition
Discrimination: Theory Two definitions are used in economic theory: narrower focus than the general issue of discrimination However, useful for a quantitative assessment Statistical Discrimination: No explicit prejudice against the minority group but asymmetric information generates differential treatment Taste Discrimination: Explicit prejudice against the minority group
Discrimination: Theory. The Taste Discrimination Model Two types of workers: Men (M) Women (W) Two types of employers: Unprejudiced (U): Worker s cost (w) = wage (w 0 ) Prejudiced (P): Worker s cost (w) if M = wage (w 0 ) Worker s cost (w) if W = wage (w 0 ) + disutility (d)
Start with the following equilibrium: W Unprejudiced W Prejudiced S S W* D D Average Wage Men = W* Q* Q Q U Q* P Average Wage Women = W* - d(q* PW /Q* W ) Question: is this equilibrium sustainable?
Possible outcome with no market failures: Perfect Segregation ASSM: types are 50/50 W Unprejudiced W Prejudiced S S W* women D D Q* W Q Q* M Average Wage Men = W* Average Wage Women = W*
Possible outcome with no market failures: Prejudiced Employers progressively eliminated from the market W Unprejudiced W Prejudiced S S U S P S W* W U women W P D D Q* U Q U Q Q P Q* P W U < W P implies Profit U > Profit P Perfect Competition implies Profit = 0 Prejudiced employers make negative profit and are eliminated
Discrimination: Empirical Evidence. Women. It is a very interesting example: widespread and common to both industrialized and developing economies Very persistent over-time Many interesting works, data, theoretical models Many results and trends can be generalized to other vulnerable groups
Empirical Evidence: Women. (cont.) Earning Differential OECD [Source: Blau and Kahn 2000 on OECD data] Country 1979-81 1989-90 1994-98 Australia 0.800 0.814 0.869 Canada 0.633 0.663 0.698 France 0.799 0.847 0.899 Germany 0.717 0.737 0.755 Japan 0.587 0.590 0.636 Sweden 0.838 0.788 0.835 UK 0.626 0.677 0.749
Empirical Evidence: Women. (cont.) Earning Differential Developing Economies [Source: World Bank 2001 on various data sources] Country 1970s 1980s 1990s Brazil 0.50 0.63 0.67 Côte d Ivoire - 0.76 - Indonesia 0.39 0.56 0.71 Kenya - 0.63 - Philippines 0.75 0.76 - Thailand - 0.74 0.80 Uganda - - 0.74
Empirical Evidence: Methodology Question: is it enough to look at raw wage differentials? No, if workers are heterogeneous Multiple regression analysis: Estimate gender wage differential conditioning on individual characteristics (human capital, demographic) That is, perform regressions of the form: log w i = x' i β + α woman + ε i i
Empirical Evidence.Women (cont.) Earning Differential US [Source: Flabbi 2006 on CPS data] Unconditional Differential.4.35.3.25.2 1980 1985 1990 1995 2000 2005 Year Conditional Differential.4.35.3.25.2 1980 1985 1990 1995 2000 2005 Year
Empirical Evidence.Women (cont.) Conditional Earning Differential US [Source: Estimates on 2005 CPS data] Woman -0.218-0.217-0.238-0.221 (0.010) (0.009) (0.008) (0.008) Control for: Demographic Yes Yes Yes Human Capital Yes Yes Industry Yes Note: reported dummy woman in a log wage regression
Empirical Evidence: Methodology Issue: we cannot control for everything. Narrower focus: decompose what we control for in Differences in characteristics (the regressors) Differences in returns on these characteristics (the estimated coefficients) Blinder-Oaxaca decomposition: logw M logw W = x' M β M x' W β W logw M logw W = ( x' x' ) β + x' ( β β ) M W M W M W
Empirical Evidence.Women (cont.) Earning Differential Decomposition A Tale of Three African Countries [Source: Appleton, Hoddinott, Krishnan 1999] Ethiopia Côte d Ivoire Uganda Total 0.243 0.053 0.331 Due to: Characteristics -0.023 0.048-0.050 Returns 0.267-0.013 0.381
Empirical Evidence.Women (cont.) There are many other important gender differentials: Education Labor market participation Access to Resources They play an important role in making a group more or less vulnerable Some take place in the labor market, some in the process that may lead to the labor market
Empirical Evidence.Women (cont.) Enrollment Rate in Primary Education [Source: World Bank 2001 on various data sources] 100 90 80 70 60 50 40 30 20 10 0 Middle East North Africa South Asia Sub-Saharan Africa Poor W Poor M All W All M
Discrimination: Solutions and Policies If a labor market issue: Equal Pay and Comparable Worth policies Affirmative Action: quota policies or pro-active policies If a pre-labor market issue: Affirmative Action in education Broader social policies
Discrimination: Solutions and Policies (cont.) Examples: Affirmative Action in Education: The US experience and the recent Supreme Court ruling The Indian experience against caste segregation Equal Pay and Comparable Worth: The role of Unions The Orchestrating Impartiality example
Segmentation: Definition Definition: similar to discrimination in the sense that there are different groups of workers that receive differential treatment in the labor market However, these groups are not generally defined by demographic characteristics Example: Unionized/Non-unionized workers Immigrant/Non-immigrant workers In the Formal/Informal sector In the Private/Public sector
Segmentation: Empirical Evidence Methodology is very similar to the one used for discrimination: compute differentials with respect to the labor market segmentation As an illustration, focus only on segmentation based on Union status: Typical example of the insider/outsider model You will see other types of segmentation such as formal/informal in other sessions of the course
Segmentation: Empirical Evidence (cont.) Nonunion/Union Wage Differential [Source: Blunch and Verner 2004; Schultz and Mwabu 1998] Ghana (1994) Conditional Conditional Bottom 10% Conditional Bottom 10% 0.014 0.156-0.051 (0.068) (0.078) (0.114) South Africa (1993) 0.895 0.468 0.107 (0.064) (0.032) (0.058) Note: reported dummy union member in a log wage regression
Segmentation: Solutions and Policies You will look at these issues in other sections of the course Briefly, they concern: Role of Unions Minimum wage Employment Protection Legislation Other policies (immigration policy, fiscal policy)
Summary Market failures in labor markets are related to the peculiarities of the good traded in them They often lead to the creation of specific institutions: some alleviate the problem other exacerbate the problem Policy interventions have a role and have been diffusely used However, inefficiencies are: still widespread and persistent both in industrialized and in developing economies
References Altonji, J.G. and R.M. Blank (1999) Race and Gender in the Labor Market, in: O. Ashenfelter and D. Card (eds.), Handbook of Labor Economics, Volume 3, Amsterdam: Elsevier Science. Appleton, S. J. Hoddinott and P. Krishnan (1999) The Gender Wage Gap in Three African Countries, Economic Development and Cultural Change, 47(2): 289-312 Blau, F. and L. Kahn (2000) Gender Differences in Pay, Journal of Economic Perspectives, 14(4): 75-99 Blau, F. and L. Kahn (2003) Understanding International Differences in the Gender Pay Gap, Journal of Labor Economics, 21(1): 106-144 Blunch N. and D. Verner (2004) Asymmetries in the Union Wage Premium in Ghana, World Bank Economic Review, 18(2): 237-252 Flabbi L (2006) Prejudice and Gender Differentials in the U.S. Labor Market in the Last Twenty Years, forthcoming, Journal of Econometrics Conference Volume: Structural Models in Labor, Aging and Health Schultz T. and G. Mwabu (1998) Labor Unions and the Distribution of Wages and Employment in South Africa, Industrial and Labor Relations Review 51(4): 680-702 World Bank (2001) Engendering Development, Oxford University Press